LinkedIn Engagement & Content Analytics

Apify actor that scrapes LinkedIn posts from company and person profile URLs and adds an engagement analytics layer on top of the same post output as the base LinkedIn post scraper. It extracts hashtags, computes an engagement score, runs sentiment analysis (lexicon or optional AI), ranks top posts, and produces an aggregate analytics summary — useful for LinkedIn content strategy, social media analytics, competitor benchmarking, marketing research, trend analysis, and B2B lead scoring.

Keywords: LinkedIn scraper, LinkedIn posts, engagement analytics, hashtag analysis, sentiment analysis, content strategy, social media analytics, competitor benchmarking, marketing insights, top posts ranking, LinkedIn data export, B2B lead quality.

What it does

• Scrapes public LinkedIn posts (JSON-LD + HTML) from company/person URLs — same engine and post schema as the base actor.
• Enriches each post with: hashtags, textLength, wordCount, engagementScore, sentiment, scrapedAt.
• Produces an analytics summary row: totals, per-post averages, top hashtags, posting-by-weekday / hour, content-type breakdown, sentiment distribution, and the top-N posts.
• Optional AI enrichment replaces the lightweight lexicon sentiment with real AI sentiment, topics, and B2B leadQuality classification across 7 providers.

Real-data policy (important)

This actor does not fabricate identifiers or counts. The base scraper emitted random tracking IDs, hashed stand-in profile/comment/reactor IDs, synthetic public IDs, fake activity URNs, and fake comment timestamps — all of that has been removed. Any field not genuinely present in LinkedIn's HTML / ld+json is returned as null. Engagement counts (numLikes, numComments, numShares) come only from each post's ld+json interactionStatistic block; when a count is unavailable it is null (never a guessed number). The analytics summary reports an honest engagementDataCoverage fraction so you know how many posts had real counts.

Input

Field	Type	Default	Description
urls	array	required	LinkedIn company or person URLs
limit	integer	10	Max posts per source URL
scrapeUntil	string	–	Keep posts on/after this date (absolute or relative)
deepScrape	boolean	true	Follow per-activity pages for more posts
rawData	boolean	false	Emit plain posts only, no analytics
topN	integer	10	Top posts / hashtags to surface
sentiment	boolean	true	Lexicon sentiment on post text
minEngagement	integer	0	Drop posts below this engagement score
metricsOnly	boolean	false	Output only the analytics summary
aiEnhancement	boolean	false	Upgrade sentiment to AI (sentiment/topics/leadQuality)
aiModel	string	claude-haiku-4-5	AI model; provider auto-detected
aiApiKey	string (secret)	–	API key for the AI provider (or env var fallback)
proxyConfiguration	object	–	Optional proxy

Output

• Dataset: enriched post rows plus one final type: "analytics" summary row (or plain posts when rawData, or only the summary when metricsOnly).
• Key-value store: linkedin_analytics.json — { "posts": [...], "analytics": {...} } (or linkedin_posts.json for rawData).

engagementScore = numLikes + 2×numComments + 3×numShares (a null count contributes 0).

Run locally

cd linkedin-post-scraper
pip install -r requirements.txt
apify run

Input example

{
  "urls": ["https://www.linkedin.com/company/google/"],
  "limit": 25,
  "scrapeUntil": "2026-01-01",
  "deepScrape": true,
  "topN": 10,
  "sentiment": true,
  "minEngagement": 0,
  "metricsOnly": false,
  "aiEnhancement": false,
  "aiModel": "claude-haiku-4-5"
}

Compliance

Scrapes only publicly available data. Respect LinkedIn's terms and applicable data-protection law when using the output.